I would like to know the simplest and surest way to set up a RGW instance with an EC-pool for storing large quantity of data.
1. I am currently trying to do this on a cluster that is not yet open to users. (i.e. I can mess around with it and, in the worst case, start all over.)
2. I deployed RGW instances using the `ceph-deploy rgw create <nodes...>`, which automatically created the following pools (the small amount of data in these pools are from a few tests I have run):
POOLS:
NAME ID USED %USED MAX AVAIL OBJECTS
.rgw.root 1 1.1 KiB 0 412 TiB 4
default.rgw.control 2 0 B 0 412 TiB 8
default.rgw.meta 3 1.2 KiB 0 412 TiB 7
default.rgw.log 4 0 B 0 412 TiB 207
default.rgw.buckets.index 5 0 B 0 412 TiB 1
default.rgw.buckets.non-ec 6 0 B 0 412 TiB 6
default.rgw.buckets.data 7 33 GiB 0 412 TiB 13564
3. These pools, by default, are all replicated (factor of 3).
4. My understanding is that the only good candidate RGW pool for EC is the .rgw.buckets.data pool, is that correct?
5. Now, given that these pools have already been created, is my only option to create a new EC-ed pool, migrate the data from the old one, and do some re-naming of the pools so that at the end, the new pool will be default.rgw.buckets.data (as outlined in https://ceph.com/geen-categorie/ceph-pool-migration/)? If so, do I have to worry about the last line of that document?: "But it does not work in all cases." (whatever that means)
6. Do I need to manually edit the CRUSH map in addition to step 5? And/or anything else I need to change or be aware of?
7. Is there an easier way to do this (i.e. create an EC-ed pool for RGW at the point of RGW initialization rather than migrating data from an old pool to a new one and renaming the new one to the old one AFTER the initial RGW pools are created) if I had not yet installed RGW? I am asking this for my future reference.
--------
Mami Hayashida
Research Computing Associate
Research Computing Infrastructure
University of Kentucky Information Technology Services
University of Kentucky Information Technology Services
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
